Octrooicentrum Nederland is het Bureau voor de Industriële Eigendom, een agentschap van het ministerie van Economische Zaken. Cognitive – ICT Patent applications in 26 fields of technology Colofon Author Jos Winnink (Patent informatics unit) DATE September 15, 2005 Ref. Nr: JWI/2005/18 STATUS Final version
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Octrooicentrum Nederland is het Bureau voor de Industriële Eigendom, een agentschap van het ministerie van Economische Zaken.
Introduction This document accompanies time series that describe the evolution of patent applications in 26 technology subfields of cognitive Information and Communication Technology (ICT). TNO asked Octrooicentrum Nederland in particular the Patinformatics Unit to extract data from patent databases and to construct the time series. The data is to be used to illustrate developments in the area of cognitive ICT Data was collected for the United States, Japan and the 25 member countries of the EU. Only documents with their oldest priority dates after 1989 were selected. From the selected documents two collections were constructed. These two collections are: 1. EP-documents
This collection consists of patent applications that were filed at the European Patent Office. The collection of EP-documents was cleaned up for documents that first were filed at the World Intellectual Property Organisation (WIPO) and at a later stage entered the EP-system.;
2. WO-documents. The WO-documents are first filed at the WIPO in Geneva;
Both collections are kept separate because of the fact that procedures for both routes differ. These differences prevent simple combination of the two collections in a way that is methodological sound. This document is accompanied by a collection of 26 spreadsheet files. There is one spreadsheet file for each technology field. All spreadsheet file contain two worksheets. The individual worksheets consist of the time series for the data from one of the two collections for every field of technology. Per worksheet there are 28 time-series. One time series for every country and one for the total counts in the technology field.. This documents ends with descriptions regarding the construction of the data for the individual technology fields.
6
Method Two databases containing patent publications were used. One is the World Patent Index (WPI) of Derwent the other is the EPODOC database of the EPO. Databases were selected in such a way that selection of documents was as precise as possible and preventing noise (unwanted documents) as much as possible. Each database system has a preferred classification system. The Derwent classification was used for WPI and the European Classification (ECLA) for EPODOC. ECLA can be considers as a more detailed version of the International Patent Classification (IPC). In most cases WPI and the Derwent classification was used. In a number of cases EPODOC with the IPC or the ECLA-system was used.. In some cases key-words had to be used in conjunction with very broad classifications like Digital Computers. In all cases the results of the selection commands were transferred tot the EPODOC-database. The reason for this is that some bibliographical information needed to create the time-series is lacking form WPI. The needed information consists of the country of residence of the applicants. In WPI related patent documents (a.k.a. families) are stored in one record whereas in EPODOC every document is stored separate. Due to this fact going from WPI to EPODOC increases the number of hits but this does not indicate the finding of new information. As mentioned in the introduction both document collections are kept separate tot prevent systematic errors. So called EURO-PCT-documents1 were removed from the EP-collection to prevent counting the same application in both collections. In the descriptions per technology field lines were typeset in boldface if the classification symbol mentioned was used in the selection. Also key words used for selecting are documented. Furthermore the selection statement and the results of the selection are documented. Assigning of a patent application to a specific country is done on the basis of the country of residence of the applicant. No corrections are made for those situations in which there is more than one applicant where these applicants are not residing in the same country. These applications are multiply counted, but their number is low and does not seriously influence the overall picture. Due to regulations patent data is published 18 or even 30 months (PCT) after application this results in the fact that the data for recent years (> 2002) is incomplete.
1 Patent applications filed at the WIPO that later enter the European regional faze are called EURO-PCT applications and also show up as separate EP-applications and are therefore counted twice of no precautions are taken.
7
Conventions
Symbolic codes The symbolic codes for the various technology fields are defined in the following table.
Technology field Symbolic code 1. artificial intelligence, AI AI 2. artificial neural network, ANN 3. bio molecular computers and artificial life BAL 4. cochlear implant, CI 5. computer vision, CV 6. expert system, ES 7. face recognition, FR 8. facial expression, FE 9. feature extraction, FEX 10. functional MRI, fMRI FMRI 11. genetic algorithm, GA 12. graphical user interface, GUI 13. human-computer interaction, HCI HCI 14. image processing, IP 15. knowledge representation, KR 16. linear discriminant analysis, LDA 17. machine learning, ML 18. machine learning technique, MLT 19. natural language processing, NLP 20. neural network model, NNM 21. pattern recognition, PR 22. principal component analysis, pca PCA 23. robotics, RO 24. speech recognition, SR 25. support vector machine, SVM SVM 26. virtual reality, VR
File names File names are constructed according the following scheme: <Symbolic code>.xls e.g. AI.xls
Worksheet names The names of the worksheets can be EP or WO. The name reflects de collection from which the time-series were constructed.
8
Country codes The country codes used are the official codes and are shown in the next table. The code tot is used for the time-series containing the data for all patent data in a technology field.
Country Code Country AT Austria BE Belgium CY Cyprus CZ Czec Republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GB United Kingdom (Great Britain) GR Greece HU Hungary IE Ireland IT Italy JP Japan LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands PL Poland PT Portugal SE Sweden SI Slovenia SK Slovakia (Slovak Republic) US United States TOT Total for the technology field
Naming convention for the time-series The naming convention for the time-series is as follows:
<symbolic code>_<collection>_<country codes> e.g. AI_EP_US
9
Technology fields Following are descriptions for the individual technology subfields: � Artificial Intelligence � Artificial Neural Network � Bio molecular computers and artificial life � Cochlear Implant � Computer Vision � Expert System � Face Recognition � Facial Expression � Feature Extraction � Functional MRI � Genetic Algorithm � Graphical User Interface � Human Computer Interaction � Image Processing � Knowledge Representation � Linear Discriminant Analysis � Machine Learning � Machine Learning Technique � Natural Language Processing � Neural Network Model � Pattern Recognition � Principal Component Analysis � Robotics � Speech Recognition � Support Vector Machine � Virtual Reality
10
1 Artificial Intelligence
1.1 Symbolic code AI
1.2 Selection codes T01 Digital Computers T01-J Data processing systems T01-J16 . Artificial intelligence (AI) T01-J16A . . Expert systems T01-J16B . . Fuzzy logic systems T01-J16C . . Knowledge processing T01-J16C1 . . . Neural networks T01-J16C2 . . . Learning T01-J16C3 . . . Natural and pictorial language processing T01-J16C4 . . . Genetic algorithms T01-J16C6 . . . Intelligent searching T01-J16C9 . . . Other AI
1.3 Corresponding IPC/ECLA codes
Derwent code IPC/ECLA code T01-J16 G06F15/18
1.4 Source database WPI
1.5 Selection statement (T01-J16 or T01-J16B or T01-J16C9)/mc
1.6 Result of selection statement 3704 hits in WPI Number of hits per individual code
1087 T01-J16 2600 T01-J16B 27 T01-J16C9
1.7 Transfer to the EPODOC database Total: 3690 documents. Time period 1990 – 2005: 3606 documents
11
2 Artificial Neural Network
2.1 Symbolic code ANN
2.2 Selection codes T01 Digital Computers T01-E Data processing T01-E05 . Novel data processing technology T01-E05B . . Neuronal configurations T01-J Data processing systems T01-J16 . Artificial intelligence (AI) T01-J16C . . Knowledge processing T01-J16C1 . . . Neural networks T02 Analogue and Hybrid Computers T02-A Analogue computers T02-A04 . Electric or magnetic computers T02-A04A . . Applications T02-A04A5 . . . Neuronal
2.5 Selection statement (T01-J16C1 or T01-E05B)/mc
2.6 Result of selection statement 4015 hits in WPI Number of hits per individual code
3885 T01-J16C1 441 T02-A04A5 not also classified using T01-E05B or T01-J16C1: 115 (These were omitted) 354 T01-E05B
2.7 Transfer to the EPODOC database Total: 8556 documents. Time period 1990 – 2005: 8416 documents
12
3 Bio molecular computers and artificial life
3.1 Symbolic code BAL
3.2 Selection codes G06N3/00 Computer systems based on biological models G06N3/00B . Bio molecular computers, i.e. using bio molecules, proteins, cells G06N3/00L . Artificial Life, i.e. computers simulating life G06N3/06 . . Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons G06N3/06B . . . using biological neurons, e.g. biological neurons connected to an integrated circuit G06N3/00 Computer systems based on biological models G06N3/12 . using genetic models G06N3/12D . . DNA computers, i.e. information processing using biological DNA
3.7 Transfer to the EPODOC database Total: 853 documents. Time period 1990 – 2005: 572 documents
13
4 Cochlear Implant
4.1 Symbolic code CI
4.2 Selection codes S05 Electrical Medical Equipment S05-F Prostheses S05-F01 . Hearing aids W04 Audio/Visual Recording and Systems W04-Y Hearing aids W04-Y05 . Characterised by type W04-Y05A . . External W04-Y05C . . Implanted W04-Y05C1 . . . With external appts. e.g. for control Keywords: implant
8.2 Selection codes T01 Digital Computers Keywords: face, facial, expression
8.3 Corresponding IPC/ECLA codes
8.4 Source database WPI
8.5 Selection statement T01/mc and ((faci+ or face+) and (expresi+ or expressi+)
8.6 Result of selection statement 605 hits in WPI
8.7 Transfer to the EPODOC database Total: 1107 documents. Time period 1990 – 2005: 1050 documents
Derwent code IPC/ECLA code
18
9 Feature Extraction
9.1 Symbolic code FEX
9.2 Selection codes G06F Electric digital data processing G06K9 Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints G06N Computer systems based on specific computational models G06T General purpose image data processing Keywords: feature, extraction
9.3 Corresponding IPC/ECLA codes
Derwent code IPC/ECLA code G06F G06K9 G06T
9.4 Source database EPODOC
9.5 Selection statement (featu+/al w extract+/al) and (g06f or g06k9 or g06t)/ec
9.6 Result of selection statement 721 hits in EPODOC
9.7 Transfer to the EPODOC database Total: 721 documents. Time period 1990 – 2005: 492 documents
13.5 Selection statement ((human w (computer or machine or robot+/al) w inter+/al) or (man-machine w interface) or (man-robot)) and (t01 or t04 or t06 or w01 or w02 or w04)/mc
13.6 Result of selection statement 424 hits in WPI
13.7 Transfer to the EPODOC database Total: 419 documents. Time period 1990 – 2005: 412 documents
23
14 Image Processing
14.1 Symbolic code IP
14.2 Selection codes S05 Electrical Medical Equipment S05-D Electrical diagnosis S05-D08 . General diagnostic processing S05-D08A . . General image processing T01 Digital Computers T01-J Data processing systems T01-J10 . For image processing T01-J10A . . Image acquisition T01-J10B . . Image processing T01-J10C . . Image generation T01-J10D . . Image digitisation/coding/compression T01-J10E . . Image storage T01-J10G . . Applications T01-J10X . . Other T04 Computer Peripheral Equipment T04-D Character and signal pattern recognition T04-D01 . Using characters containing code marks T04-D02 . Image acquisition T04-D03 . Image preprocessing for image recognition T04-D03A . . Noise reduction T04-D03B . . Edge recognition and determining orientation T04-D04 . Recognition T04-D05 . Monitoring and error detection T04-D07 . Applications of recognition techniques T04-D08 . Colour systems T04-D09 . Other recognition aspects W04 Audio/Visual Recording and Systems W04-M Video and synchronising signal generators W04-M01 . Video cameras W04-M01A . . Camera tube arrangements W04-M01B . . Solid state pick-up device arrangements W04-M01C . . (Auto)focusing, zooming, lenses for TV camera, shutters, filters W04-M01D . . Control circuits, monitoring, displays, viewfinders W04-M01D6 . . . Image processing and function control W04-M01D6A . . . . Image acquisition aspects W04-M01L . . Stereoscopic image generating camera system W04-M09 . Other video source aspects
18.5 Selection statement (machin+/al w learning+/al w techniq+/al)
18.6 Result of selection statement 35 hits in EPODOC
18.7 Transfer to the EPODOC database Total: 35 documents. Time period 1990 – 2005: 35 documents
29
19 Natural Language Processing
19.1 Symbolic code NLP
19.2 Selection codes G06F17/00 Digital computing or data processing equipment or methods, specially adapted for specific functions G06F17/20 . Handling natural language data G06F17/21 . . Text processing G06F17/22 . . . Manipulating or registering by use of codes, e.g. in sequence of text characters G06F17/24 . . . Editing G06F17/25 . . . Automatic justification G06F17/26 . . . Automatic hyphenation G06F17/27 . . Automatic analysis, e.g. G06F17/28 . . Processing or translating of natural language G06F17/28D . . . Data Driven translation
19.3 Corresponding IPC/ECLA codes
Derwent code IPC/ECLA code G06F17/20 – G06F17/28D
19.4 Source database EPODOC
19.5 Selection statement G06F17/2+/al/ec
19.6 Result of selection statement 17361 hits in EPODOC
19.7 Transfer to the EPODOC database Total: 17361 documents. Time period 1990 – 2005: 9873 documents
30
20 Neural Network Model
20.1 Symbolic code NNM
20.2 Selection codes G06N3/00 Computer systems based on biological models G06N3/02 . using neural network models G06N3/04 . . Architectures, e.g. interconnection topology G06N3/06 . . Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
20.5 Selection statement (G06N3/02+ or G06N3/04+ or G06N3/06+)/ec
20.6 Result of selection statement 9811 hits in EPODOC Number of hits per individual code
1113 G06N3/02+ 5320 G06N3/04+ 3805 G06N3/06+
20.7 Transfer to the EPODOC database Total: 9811 documents. Time period 1990 – 2005: 2547 documents
31
21 Pattern Recognition
21.1 Symbolic code PR
21.2 Selection codes T04 Computer Peripheral Equipment T04-D Character and signal pattern recognition T04-D01 . Using characters containing code marks T04-D02 . Image acquisition T04-D02A . . Mechanical and optical aspects of image acquisition T04-D02B . . Circuitry, processing of image acquisition T04-D03 . Image preprocessing for image recognition T04-D03A . . Noise reduction T04-D03B . . Edge recognition and determining orientation T04-D04 . Recognition T04-D05 . Monitoring and error detection T04-D07 . Applications of recognition techniques T04-D07A . . Detecting defect in pattern T04-D07B . . Sorting objects by type T04-D07B1 . . . Using patterns specifically applied as identification marks T04-D07C . . Identification of item T04-D07D . . Detecting movement or position T04-D07D1 . . . Detecting movement T04-D07D5 . . . Detecting position or orientation T04-D07E . . Hand written character recognition T04-D07K . . Using non-visible light images (e.g. IR,UV) T04-D07X . . Other recognition applications T04-D08 . Colour systems T04-D09 . Other recognition aspects T07 Traffic Control Systems T07-A Determining road vehicle position, speed or flow T07-A03 . Identifying and recording individual vehicle information T07-A03C . . Recording images T07-A03C5 . . . By video systems T07-A03C5A . . . . With pattern recognition of licence plate information
22.5 Selection statement (principal w component w analysis) and (t01 or t04)/mc
22.6 Result of selection statement 119 hits in WPI
22.7 Transfer to the EPODOC database Total: 274 documents. Time period 1990 – 2005: 264 documents
34
23 Robotics
23.1 Symbolic code RO
23.2 Selection codes T01 Digital Computers T01-J Data processing systems T01-J07 . For industrial process control T01-J07B . . Computer control of manufacturing/industrial machine and quality control T01-J07B1 . . . Quality control T01-J07B2 . . . Semiconductor manufacture control V03 Switches, Relays V03-U Switches/relays characterised by applications V03-U14 . Robotics V04 Printed Circuits and Connectors V04-M Connectors for specific applications V04-M30 . Characterised by application to specific industry V04-M30R . . Machine tools; robotics V04 Printed Circuits and Connectors V04-Q Printed circuits V04-Q30 . Characterised by application to specific industry or equipment V04-Q30R . . Machine tools; robotics V06 Electromechanical Transducers and Small Machines V06-U Electric machines characterised by applications V06-U05 . Robotic